Google is in trouble! The tech giant couldn’t seem to come up with a good solution to fixing its photo app after it categorized a Black couple as “gorillas” in 2015.
Instead of properly fixing its image recognition software algorithm recently, it only banned the word “gorillas” from its auto-tag tool, the Daily Mail reported. Google only blocked identification of gorillas, chimpanzees and monkeys, a half-baked measure that does not effectively address the problem of its racist process for labeling photos.
The company’s attempt to address the problem was slammed as lame and problematic. Google missed the point that a diverse model for its algorithm is needed, not a quick “fix,” several social media users said.
Google launched its photos app in May 2015, creating image collections of people and landscapes. Its auto-tagging feature was designed to organize uploaded photos and make searching easier. However, the tech giant pissed a lot of people off when one New York couple found pictures of themselves under a collection with a racist name. Jacky Alcine, of Brooklyn, New York, spotted photos of him and a female friend grouped into a collection tagged “gorilla” shortly after the app’s debut.
The tech giant was ‘appalled’ and ‘genuinely sorry’ for the mistake, the company said after the racist incident. “Image labeling technology is still early and unfortunately it’s nowhere near perfect,” a company spokesperson said.
A diverse app would have prevented the problem, one professor said online.
“How about making the effort to develop a diverse (eg not all white) model base for the algorithm?’ tweeted Agustin Fuentes, a US-based Professor of Anthropology.
If Google had more employees of color, the problem could have been avoided. The number of Black U.S. employees remained at 1 percent of tech workers and 2 percent of the total workforce in 2017, stagnant over the past year, according to Google’s EEO-1 report cited by Fortune.